Update: Sensor-Augmented Insulin Pumps
The New England Journal of Medicine was the first to publish the Diabetes Control and Complications Trial in 1993. Understanding the importance of research in diabetes, the journal continues to publish the latest evidence-based literature to achieve both a “cure” and state of the art “care” of diabetes. Following my most recent blog about insulin pumps and multiple daily injections, the paper “Threshold-Based Insulin-Pump Interruption for Reduction of hypoglycemia” by Bergenstal, Klonoff, Garg, Bode, Meredith, Solver, Ahmann, Welsh, Lee and Kaufman (ASPIRE study group) was published in the journal in the July 18 issue (N Engl J Med 2013; 369:224-32.).
The present ultimate goal in the care of diabetes is develop an “artificial pancreas” in which the glucose sensor communicates with the insulin pump to affect the delivery of insulin. Key to the safety of the eventual artificial pancreas is the ability to prevent an overdose of insulin and thus a severe hypoglycemic episode. Much effort has been directed toward the to development of an algorithm to suspend the pump when the blood sugar has reached a pre-programmed level.
The ASPIRE in-home study group (Automation to Stimulate Pancreatic Insulin Response) has been working to “marry” the pump to the sensor as the first crucial step in the development of the artificial pancreas. Thus, the “threshold-suspend feature of these sensor-augmented pumps is meant to “minimize the risk of hypoglycemia by interrupting insulin delivery at a preset sensor glucose value.” The authors evaluated the sensor-augmented insulin pump therapy by comparing patients with and without the threshold-suspend feature in patients with nighttime hypoglycemia.
Who were the participants in the study?
Patients that were eligible for the study included those between 16 and 70 years of age and had type 1 diabetes for at least two years. Other criteria included a hb A1c of 5.8-10.0% and use of insulin pump therapy for over six months. Two hundred forty seven (247) patients with type 1 diabetes and documented nighttime hypoglycemia were randomly assigned to receive sensor-augmented insulin pump therapy with (121 patients, the experimental group) or without (126 patients, the control group) the threshold-suspend feature over a –three month period. There were two main objectives of the study: a safety outcome and an efficacy outcome (success of the intervention)
- safety outcome: the change in the hb A1c
- efficacy outcome: area under the curve for nighttime hypoglycemic events (an indirect means to measure the severity and duration of studied events)
Two hour threshold-suspend events were studied in relation to subsequent sensor glucose values.
What are the results of the study?
- The changes in the hb A1c values were similar in the two groups.
- The mean area under the curve (the severity and duration of nighttime hypoglycemia) was 37.5% lower in the threshold-suspend group than in the control group.
- Nighttime hypoglycemia occurred 31.8% less frequently in the threshold-suspend group than in the control group.
- The percentages of nocturnal sensor glucose values <50 mg/dl, 50-<60 mg/dl, 60-<70 mg/dl were significantly reduced in the threshold-suspend group (P<0.001 for each range).
- There were 1438 suspensions during the night in which the pump was stopped for 2 hours resulting in an average sensor glucose value of 92.6 mg/dl.
- Four patients (all in the control group) had a severe hypoglycemic episode and there were no instances of DKA.
What may we conclude from this study?
The application of a sensor-augmented insulin pump with a threshold –suspend feature decreased nighttime hypoglycemia without increasing hb A1c levels. Interestingly enough, the four episodes of severe hypoglycemia that occurred in the control group during the study were “identical to the rate among patients that receive multiple daily injections and those who received sensor-augmented pump therapy” in a previous study (without the suspend feature.)
What are the limitations of this study?
- There was no validation of sensor glucose values in the analyses by another reference method.
- Blood glucose sensor values of <65 mg/dl lasting <20 minutes and those with evidence of pump interaction were not analyzed.
- Only those patients that were prone to hypoglycemia were enrolled; thus perhaps decreasing the generalizability of the study.
- Three month duration of the study may not have been long enough to demonstrate an improvement to quality of life.
Based on the positive results of this study, it appears that we remain on the path to the successful creation of the artificial pancreas.
I will continue to keep you posted.